Technical Note

Rapid and massive throughput analysis of a constant volumehigh-pressure gas injection system

Xiaoli Ren a, b, c, Jia Zhai b, *, Jihong Wang a, b, Ge Ren a, b

a Key Laboratory of Optical Engineering, Chinese Academy of Sciences, Chengdu, 610209, Chinab Institute of Optics and Electronics, Chinese Academy of Sciences, Chengdu, 610209, Chinac University of Chinese Academy of Sciences, Beijing, 100049, China

a r t i c l e i n f o

Article history:Received 24 July 2018Received in revised form19 November 2018Accepted 3 December 2018Available online 5 December 2018

Keywords:FPSSGas injectionTransient experimentFlow developmentPrescribed time

a b s t r a c t

Fusion power shutdown system (FPSS) is a safety system to stop plasma in case of accidents or incidents.The gas injection system for the FPSS presented in this work is designed to research the flow develop-ment in a closed system. As the efficiency of the system is a crucial property, plenty of experiments areexecuted to get optimum parameters. In this system, the flow is driven by the pressure difference be-tween a gas storage tank and a vacuum vessel with a source pressure. The idea is based on a constantvolume system without extra source gases to guarantee rapid response and high throughput. Amongthem, valves and gas species are studied because their properties could influence the velocity of the fluidfield. Then source pressures and volumes are emphasized to investigate the volume flow rate of theinjection. The source pressure has a considerable effect on the injected volume. From the data, properparameters are extracted to achieve the best performance of the FPSS. Finally, experimental results areused as a quantitative benchmark for simulations which can add our understanding of the inner gas flowin the pipeline. In generally, there is a good consistency and the obtained correlations will be applied infurther study and design for the FPSS.© 2018 Korean Nuclear Society, Published by Elsevier Korea LLC. This is an open access article under the

CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction

The international thermonuclear experimental reactor (ITER) isa fusion reactor, in which the plasma is contained in a doughnut-shaped vacuum vessel. During plasma discharge, the fuel - amixture of Deuterium and Tritium, is heated to temperatures inexcess of 150 million �C. A hot plasma is formed which needs to beremoved by cooling water system. In some accidents, such as thecooling water pipe break, the high temperature water would leakout and the cooling capability will become insufficient. Without thetermination system for plasma, the heat load will cause melting ofcomponents which could produce hydrogen. So the mixed air couldcreate an explosion which will exceed the design limit of the vac-uum vessel (VV).

Fusion power shutdown system (FPSS) is a gas injection systemwhich is required to terminate plasma in accidental events. TwoFPSS facilities can be executed in parallel. The simplest structure ofthis system consists of a reservoir of gas and a dedicated tube for

gas injection. As a safety device, quite a few special apparatusrequire gas injection systems to inject high throughput gas in shorttimes without air supply. For example, Tokamak is injected~4 � 1022 particles (atoms or molecules) within 2e5 ms to provideadequate gas tomitigate the disruption-caused damage. An injectorshould provide the plasma with larger amounts of neon to radiatemost of the thermal energy, and force the current decay time towithin a 50e150 ms interval [1]. Theoretically, components of thesystem (such as valves, pipes, etc.) and process parameters (such asflow rate, pressure difference, etc.) have great influence on the in-jection system [2]. Therefore, it is necessary to investigate theperformance of the flow field influenced by components andoperation processes [3].

In fact, continuous gas supply can achieve fast response andhigh gas injection. Xiuquan Cao used a mass flow controller(maximum flow rate 25 l/min) to control the working gas suppliedto the plasmawith an accurate and high gas flow rate [4]. S. C. Batesobtained up to 500 Torr liter/s for hydrogen from a fast valve with3 ms [5]. Min Jae Kim conducted a high mass flow rate gas inlet of89.36 kg/s to guarantee high output to the air storage system [6]. Inthe work of Toufik Boushaki, using a constant flow rate for all* Corresponding author.

E-mail address: un31415926@163.com (J. Zhai).

Contents lists available at ScienceDirect

Nuclear Engineering and Technology

journal homepage: www.elsevier .com/locate/net

https://doi.org/10.1016/j.net.2018.12.0011738-5733/© 2018 Korean Nuclear Society, Published by Elsevier Korea LLC. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Nuclear Engineering and Technology 51 (2019) 908e914

http://creativecommons.org/licenses/by-nc-nd/4.0/mailto:un31415926@163.comhttp://crossmark.crossref.org/dialog/?doi=10.1016/j.net.2018.12.001&domain=pdfwww.sciencedirect.com/science/journal/17385733www.elsevier.com/locate/nethttps://doi.org/10.1016/j.net.2018.12.001http://creativecommons.org/licenses/by-nc-nd/4.0/http://creativecommons.org/licenses/by-nc-nd/4.0/https://doi.org/10.1016/j.net.2018.12.001https://doi.org/10.1016/j.net.2018.12.001

experiments to ensure high velocity and high quantity introducedgas [7]. This gas supply needs extra source gases and tubes whichcould increase the response time of the injection system.

Application characteristics of all components used in the in-jection system had been discussed. Ki-Yong Choi used a quickopening valve which was triggered within 0.5 s to simulate a directvessel injection line [8]. When the sheath gas was injected throughthe normal nozzle, the flow rate is higher than that of the othernozzles [9]. The time of injection and the time between successiveinjections could be controlled by a pressure regulation valvebecause the pressure of injection was manually controlled with it[10]. Eddy current actuated valve could be performed to injectmassive gas to plasma in 1e2 ms [11]. From Keith H. Burrell, thesub-millimeter diameter tubing and piezoelectric gas puff valveswere used in the system to obtain a fast response time [12e15].Nevertheless, all these factors can decide the property of the systemand they affect mutually. Thus studying a factor solely cannotreflect the total phenomenon of the FPSS, such as velocity, density,response time, etc.

In this work, we present a gas injection system to satisfy theFPSS requirements. All the factors are considered according to theactual application for ITER, including constant volume system,rapid response time, high pressure difference, high throughput andbig flow resistance. Then simulation models are analyzed to verifythe reliability of the experiment and numerical data can assist toaccount for the internal flow filed in this system. All the data of thiswork are to experimentally observe the specific phenomena of thesystem without gas supply. And these results can be used forfurther studies and designs of the FPSS.

2. Experiments

Based on the massive gas injection concept and the plasmashutdown time, the FPSS could inject gas over a minimal volume of3000 Pa.m3 with a limited time of 3 s. A tradeoff exists betweentime response, source pressure, source volume and injected vol-ume. Details of the experimental apparatus and procedure havebeen published detailedly here.

2.1. Experimental apparatus and procedure

Fig. 1 depicts the schematic of experimental apparatus. Thesystem has some major components: gas storage tank, flow pipe,vacuum vessel, valve, pressure transducer and vacuum pump.Through the closed pipe circuit, the upstream has being controlledby valves and monitored by manometers. During the process, afterthe valve is open, the gas flows from the tank to the vessel throughthe long pipe. Then the vessel pressure could be obtained to deduce

the injected gas while the valve is closed at 3 s.The model of the gas pipe shown in this manuscript was

designed at a one-to-one scale of ITER. The pipe has length, L, of23.4 m and inner diameter, D, of 0.01022 m. The valve is installed inthe outlet of the tank. Manometers (MKS Inc. USA) are separatelylocated at some positions for the purposes of valve feedback controland injected volume judgment. Among them, one is fitted on thevessel to monitor its pressure, one is fixed at the inlet of the vesselto determine the arrival time of the gas flow, and others aremounted on the inlet and outlet of the valve near the gas supply togather source pressures. The reading accuracy of all the abovemanometers is 0.5%. The vacuum vessel has volume, Vv, of 4.5 m3. A0.2 m3/s roots pump is used to evacuate the vacuum vessel pres-sure, Pv, of 0.01 Pa. In the experiment, the vacuum chamber keepsthe vacuum at the initial time and the vacuum pump does not workduring the gas injection process. All the components are made ofstainless steel and the experimental model can be remotelycontrolled and calibrated.

2.2. Response times

The flow rate adjusts itself to give a pressure drop equaled todifferences between entrance and exit pressures. The responsetime for flow is a function of average flow velocity, density, vis-cosity, pressure and pipe length, diameter and other accessories.Nevertheless, the structure of the long pipe is irregular as it islimited by the practical installation condition. The length range ofeach segment is from 51 mm to 3676 mm, and the angle range ofeach bend is from 46� to 175�. There are 26 bends, and all thebending radius are 40mm. Those dramatic changes in structure canyield choked flow due to the friction effects (Fig. 2). So there is astable response time of the pipe which cannot be modified artifi-cially. Except this qualification, response times varied with valves,gases, volumes and pressures could be developed.

2.2.1. Fast-acting valve response timesIn this gas injection system, a controlled explosion creates a

brief flow from a gas storage tank with data taken by fast-responsevalves. The whole period of the transient flow considered in thestudy is unknown. Generally, response times of valves are some-what dependent upon the particular manufacturer’s design. How-ever, the specific opening and closure times, especially for valvesdriven by pneumatic actuators, must be pre-confirmed. Thepneumatic valve chosen for this application are VAT all-metal anglevalve and Swagelok series 1 valve [16,17]. Table 1 lists the specifi-cations of them.

In the test method, the computer sends a command to open thevalve, meanwhile, it sends out a trigger pulse to the oscilloscope.Then the pressure graphics can be amplified and read out on theoscilloscope. According to the pressure sensor signal, the responsetime of the valve in this experiment can be assured. The electronicdeveloped to control the pneumatic actuator applies a maximumvoltage of 24 V at the onset of demand. This voltage opens the valve

Inert gas Isolation valve Tank Manometer 1

Pneumatic valve Manometer 2 Manometer 3 Manometer 4

Vacuum vessel Pump Compressed air Manometer 5

Fig. 1. Diagram of the gas injection system.

Length range: 51 mm - 3676 mm

Angle range: 46° - 175°

Bending radius: 40 mm

Fig. 2. Layout of the pipeline.

X. Ren et al. / Nuclear Engineering and Technology 51 (2019) 908e914 909

actuator in the shortest time. Under certain conditions, oscillationsare observed when the valve is opened and we studied the rise andfall times of the flow.

Three different processes consist the pressure curve (Fig. 3). Thefirst process is the initial stage of constant acceleration, wherein thecompressed air propagates from the electromagnetic device. Thenthe pneumatic valve is opened by the actuator, so the pressureincreases under the initial high pressures difference. The secondprocess is the subsequent stage of constant pressure, wherein thegas flows from the tank into the pipe through the pneumatic valve.The pressure maintains stably since the flow field reaches dynamicequilibrium. The final stage is the closing process of constantdeceleration, wherein the electromagnetic device outputs lowvoltage to stop the control pressure until the pneumatic valve isclosed. The pressure immediately decreases because of the termi-nated gas source.

As shown in Fig. 4 (a), the feedback voltage is received from thepneumatic valve. The response of Swagelok valve is faster than thatof VAT valve and specific parameters can be obtained from Fig. 4 (b).The minimal response time of the VAT is about 140 ms, but themaximal response time of the Swagelok valve is only 60 ms. So theSwagelok valve is the optimum selection for this work.

The compressed air is used to open the pneumatic valve. Theamount of gas consumed when the actuator is working can beconfirmed as [18].

Q ¼ V� 103 � ½ðPþ 0:1033Þ=0:1033� (1.1)

Where Q is the consumed gas amount, V is the volume of theactuator, P is the control pressure.

To a pneumatic valve, its volume of the actuator is specified.When the control pressure is larger, the consumed gas can quicklyreach the required amount. Additional experiments are desirable tofurther understand this relationship. The control pressure ismonitored at the interface between the compressed air and theactuator. For very small control pressure difference, however, theforward gain is not enough to provide the maximum dynamic. Sothe valve takes a little longer time to open (about a few millisec-onds). From experiments, the fast response occurs on the maximalcontrol pressure of 0.7 MPa and a quick response time of 50 ms hasbeen shown (Fig. 5).

2.2.2. Test gas response timeIn general gas injection systems, the response time of the gas

species has always been ignored. But in this paper, this index has tobe investigated because it could affect the whole time of the flowcircuit.

In stationary isentropic flows and in the absence of externalforces, the Bernoulli’s theorem can be written as [19].

Table 1Specifications of two valves.

Characteristics VAT angle valve Swagelok series 1

Inner Diameter (mm) 10 7.6Control Pressure (MPa) 0.6e0.8 0.5e0.7Overpressure (MPa) 0.6e0.9 0.3e1.03Feedback Voltage (V) Closure 0 6

Opening 6 0

Fig. 3. The trigger signal.

)b()a(Fig. 4. Valve response times.

X. Ren et al. / Nuclear Engineering and Technology 51 (2019) 908e914910

Dp ¼ �ru2

2þ constant (1.2)

Where Dp is Pressure drop, r is gas density, u is mean velocity.This solution suggests that if the velocity in channel flow with

constant pressure gradient increases, the gas density mustdecrease. Inert gas released from the FPSS is used to prevent fusionreaction and test gases are nitrogen (Ne) and argon (Ar). Amongthem, the density of the Ne is 0.9002 kg/m3, and the Ar is 1.7818 kg/m3. Based on the Bernoulli’s theorem, the Ne is appropriate for thissystem.

Table 2 depicts that the response time of the Ne is alwaysshorter than that of the Ar under different source pressures. Thisconsequence has great agreement with the theory.

2.3. Gas injection volumes

2.3.1. Pipe flow equationsAnother important index in this study is the injected inert gas of

3000 Pa.m3 to the vacuum vessel. This volume is deduced by thevessel pressure which is calibrated by a capacitance manometer.Therefore, parameters of the gas source which could decide thethroughput must be confirmed, such as the source volume and thepressure.

In the pipe flow, the pressure dropped down the pipe is relatedto pipe parameters. It can be expressed as

Dp ¼ 8hula2

(1.3)

Where h is coefficient of viscosity, l is tube length and ɑ is tuberadius.

The gas expands and accelerates into the vacuum vessel at theend of the pipe. Independent of the fluid state in the tube, thepressure dropped to vacuum is always greater than it in other po-sitions. So the end of the tube acts as a sonic nozzle. The volumeflow rate is

Q ¼ Dp*pa4

8hl(1.4)

This equation shows the relation between the volume flow rateand the pressure drop.

Fig. 5. The response time of the Swagelok valve.

Table 2Response times of Ne and Ar gases.

Source Pressure (MPa) Response Times (ms)

Ne Ar

0.2 132 1340.3 115 1200.4 111 1170.5 109 1140.6 106 1121 98 110

Manometer 2Manometer 1

Swagelok

Tank

Compressed air

Manometer 3

Stop valve Vacuum vesselManometer 4

Pump

Fig. 6. Experimental layout.

Fig. 7. Experimental data for source pressure.

X. Ren et al. / Nuclear Engineering and Technology 51 (2019) 908e914 911

2.3.2. Gas source pressureTo ensure that the gas source is sufficient to produce meaningful

injected volumes, the source pressure must be confirmed firstly. Inthis experiment, the test volume of the gas storage tank is 0.001 m3,themeasured sourcepressure is increased from0.1MPa to1MPa, andthe Swagelok series 1 valve is chosen (Fig. 6). The stop valve is closedand quickly sealed after the gas is injected into the vacuum vessel in3s. Two pressure manometers are located on two ends of the valve.

Fig. 7 displays the analog signal showed in the oscilloscope andthe pressure change obtained from manometers. After the valve isopened, manometer 1 indicates a decay time of 0.2 s. This timeincludes the gas flow time and the response time of the valve andthe manometer. When the neon gas exits the tank and is excited bythe initially high pressure difference, the sudden increase data ofthe manometer 2 appears.

The data detected from manometer 1 and 3 are shown in Fig. 8as the flow rate time reaches 3s. The manometer 1 collects tankpressures which decrease rapidly while the valve is opened. Untilthe pressure reaches about 100 kPa, the descending trend begins tobecome slow. The reason is that the pressure difference is verysmall between the tank and the vessel at this time. The manometer3 gathers the pressure at the end of the pipe (inlet of the vacuumvessel). As the gas flow reaches the vacuum vessel after 0.3 s, thepressure increases largely. After the value reached the peak, thepressuremaintains a high level and decelerates slowly. It elucidatedthat there is still high pressure difference at both ends of themanometer. All the curves above show that the accelerates of therise velocity and the peak value increases with larger pressuredifference.

Following a gas pulse, the total gas efflux is obtained by an in-dependent measurement 4 installed on the vacuum vessel. During3s, the injection volume of the Ne is significantly larger than that ofAr. The profile represents well the injected volume evolution of theflow in the vacuum vessel (Fig. 9). This figure also reveals that thesystem can achieve injection volume to 3000 Pa.m3 with a sourcepressure of 3.3 MPa. However, this source pressure exceeds theworking range of the Swagelok valve, hence a suitable source vol-ume must be insured.

2.3.3. Gas source volumeFive gas source volumes are considered in this work, including

0.001m3, 0.002m3, 0.004m3, 0.006m3 and 0.04m3. Two Swagelokdouble-ended cylinders of 0.002 m3 is connected in series to form atank of 0.004 m3, and this is also true for the tank of 0.006 m3.

Fig. 10 shows results of the injected gas volume in the vacuumvessel. Under the maximal working pressure of the Swagelok

)b()a(

Fig. 8. Experimental data (a) Tank pressures; (b) Pressures at the inlet of the vacuum vessel.

Fig. 9. Injected volume in the vacuum vessel with two gas species.

Fig. 10. Injected volume in the vacuum vessel with difference source volume.

X. Ren et al. / Nuclear Engineering and Technology 51 (2019) 908e914912

valve, the source volume must be lager than 0.005 m3, only thenit may inject adquate gas volume. Nevertheless, the innerdiameter of the interface between two tanks is much small. Thiscontraction structure could add extra resistance to the flow field,so the system response time may be increased. Meanwhile, thetandem cylinders also ask more installing space, which is notrealizable in the FPSS.

Fortunately, another new type of Swagelok valve is provided forthe work, which is described as series 5. This kind of the valve hasthe identical performance with series 1, and its maximal workingpressure could reach 6.89 MPa.

Under the source volume of 0.001 m3, experiments are carriedout to ensure the measured data can satisfied the system

requirement. Notice that for almost all the data, the injected vol-ume versus pressure relation is basically linear. Fig. 11 presentsthat under the identical source pressure, two series of valves caninject the same gas volume and the required source pressure isabout 3.26 MPa.

Through above experiments, some parameters of best fit in theFPSS are given in Table 3.

3. Simulation verification

We employed CFD (Computational Fluid Dynamics) to demon-strate and understand inner flow phenomenon. The model andboundary conditions are uniformwith experimental parameters. Inthe computational procedure, a segregated implicit solver andsecond-order upwind scheme are employed. A time-step ofDt ¼ 1� 10�5 s is applied to achieve convergence at each iteration.This convergence of the computed solution is determined based onresiduals set at 10�3 for the mass conservation and momentumequations [20].

As shown in Fig. 12, all the data have similar tendencies;nevertheless, computational values are less than experiment re-sults. These discrepancies can be due to either inadequate storageor inability to simulate the turbulent flow of irregular geometries.

To generate sufficient injected gas quantity for the vacuum

Fig. 11. Injected volume in the vacuum vessel with two Swagelok valves.

Table 3Optimum parameters for the gas injection system.

Characteristics Parameters

Pneumatic valve Swagelok series 5control pressure 0.7 MPaInert gas NeGas source volume 0.001 m3

Gas source pressure 3.3 MPa

Fig. 12. Comparison on experimental data and simulation results.

Fig. 13. Relationship between the injected volume and the source pressure.

X. Ren et al. / Nuclear Engineering and Technology 51 (2019) 908e914 913

vessel, a higher source pressure is needed with 3.12 MPa in Fig. 13.The total gas injection volume obtained through experiments andsimulations are compared with each other. A good agreement isfound with the both methods.

4. Conclusions

This paper demonstrates a gas injection system for the FPSSwhich is a constant volume system and has not been supplied extrasource gas. Meantime, the pipe length is predetermined and itsarrangement is irregular. But this system could satisfy the fastresponse and high throughput demands. Experimental in-vestigations of the effect of pressure, volume, gas type, tube length,and valve on the requirement have been presented. They aremeasured in accordance with the design specification for thepractical necessary. The resulting data adds our overall under-standing of how strong response times, source pressures and vol-ume influence the flow development.

In this study, the neon throughput of up to 3000 Pa.m3 could beadequate for vacuum vessel to stop the hazard fusion reaction.Through comparisons to simulation investigations, the volume ofinjected gas is estimated from the pressure difference in the vesselbefore and after the gas injection. Results show that the sourcepressure for the gas storage tank must be above 3.3 MPa. Simul-taneously, injected gas is linear scale to the source pressure and thesimulation also displays a similar tendency. So CFD codes for in-jection systems are validated and the simulation phenomenoncould enhance our understanding of the internal fluid flow in theFPSS.

The device has recently been installed in the practical engi-neering application of the FPSS and first data has been obtained.Based on this gas injection system, the next step includes studyingthe components distribution and the structure design of the gassource, the trigger mechanism, the control and monitoring methodof the FPSS. Meanwhile, more comprehensive experiments will beperformed to verify those researches.

Disclaimer

The views and opinions expressed herein do not necessarilyreflect those of the ITER organization.

References

[1] G. Pautasso, M. Bernert, M. Dibon, B. Duval, R. Dux, E. Fable, J.C. Fuchs,

G.D. Conway, L. Giannone, A. Gude, A. Herrmann, M. Hoelzl, P.J. McCarthy,A. Mlynek, M. Maraschek, E. Nardon, G. Papp, S. Potzel, C. Rapson,B. Sieglin, W. Suttrop, W. Treutterer, The ASDEX Upgrade team and theEURO fusion MST1 team. Disruption mitigation by injection of smallquantities of noble gas in ASDEX upgrade, Plasma Phys. Contr. Fusion 59(2017) 1e11.

[2] R.A. Langley, J.W. Halliwell, G.R. Dyer, J.L. Yarber, Gas injection system for theadvanced toroidal facility, J. Vac. Sci. Technol. 7 (3) (1989) 2423e2426.

[3] D.G. Whyte, T.C. Jernigan, D.A. Humphreys, A.W. Hyatt, C.J. Lasnier, P.B. Parks,T.E. Evans, M.N. Rosenbluth, P.L. Taylor, A.G. Kellman, D.S. Gray,E.M. Hollmann, S.K. Combs, Mitigation of Tokamak disruptions using high-pressure gas injection, Physical Review Letters 89 (5) (2002) 1e4.

[4] Xiuquan Cao, Deping Yu, Yong Xiang, Jin Yao, Influence of the gas injectionangle on the jet characteristics of a non-transferred DC plasma torch, PlasmaChem. Plasma Process. (36) (2016) 881e889.

[5] S.C. Bates, K.H. Burrell, Fast hydrogen gas injection system for plasma physicsexperiments, Rev. Sci. Instrum. 55 (6) (1984) 934e993.

[6] Min Jae Kim, Tong Seop Kim, Feasibility study on the influence of steam in-jection in the compressed air energy storage system, Energy (141) (2017)239e249.

[7] Toufik Boushaki, Jean-Charles Sautet, Characteristics of flow from an oxy-fuelburner with separated jets: influence of jet injection angle, Exp. Fluid (48)(2010) 1095e1108.

[8] Ki-Yong Choi, Hyun-Sik Park, Seok Cho, Kyoung-Ho Kang, Nam-Hyun Shoi,Dae-Hun Kim, Choon-Kyung Park, Yeon-Sik Kim, Won-Pil Baek, Experimentalsimulation of a direct vessel injection line break of the Apr1400 with the atlas,Nuclear Engineering and Technology 41 (5) (2009) 655e676.

[9] Vadikkeettil1 Yugesh, Ravi Ganesh, Ramachandran Kandasamy, Vidhi Goyal,Kailsha Chandra Meher, Influence of the shroud gas injection configuration onthe characteristics of a DC non-transferred arc plasma torch, Plasma Chem.Plasma Process. 38 (4) (2018) 759e770.

[10] M.R.O. Panao, A.L.N. Moreira, Flow characteristics of spray impingement in PFIinjection systems, Exp. Fluid (39) (2005) 364e374.

[11] A.J. Thornton, K.J. Gibson, J.R. Harrison, M. Lehnen, R. Martin, A. Kirk, the MASTTeam, Characterization of disruption mitigation via massive gas injection onMAST, Plasma Phys. Contr. Fusion 54 (2012) 1e14.

[12] K.H. Burrell, Fast gas Injection system for plasma physics experiments, Rev.Sci. Instrum. 55 (6) (1984) 934e993.

[13] A. Beune, J.G.M. Kuerten, M.P.C. van Heumen, CFD analysis withfluidestructure interaction of opening high-pressure safety valves, Comput.Fluids (64) (2012) 108e116.

[14] Adrian Constantin Clenci, Victor Iorga-Siman, Michael Deligant,Podevin Pierre, Georges Descombes, Rodica Niculescu, A CFD (computationalfluid dynamics) study on the effects of operating an engine with low intakevalve lift at idle corresponding speed, Energy (4) (2014) 1e16.

[15] Craig Warren Gustafson, The Fluid Mechanics of Hydraulic Fracturing, Uni-versity of Illinois at Urbana Champaign, 1987, p. 322.

[16] SwagelokdTM Swagelok Company. Sample Cylinders, Accessories, andOutage Tubes[Z], Swagelok Company, U.S.A, 2013.

[17] VAT, Inc, Vat Vacuum Valves 2016[Z], VAT Inc, U.S.A, 2016.[18] Binghui Xu, Pneumatic manual[M], Shanghai Science and Technology Press,

China, 2005.[19] Frank M. White, Fluid Mechanics[M], University of Rhode Island, New York,

1997.[20] Xiaoli Ren, Jia Zhai, Ge Ren, Numerical study of transient flow in the pre-

liminary design of fusion power shutdown system, Fusion Eng. Des. 129(2018) 214e220.

X. Ren et al. / Nuclear Engineering and Technology 51 (2019) 908e914914

http://refhub.elsevier.com/S1738-5733(18)30540-0/sref1http://refhub.elsevier.com/S1738-5733(18)30540-0/sref1http://refhub.elsevier.com/S1738-5733(18)30540-0/sref1http://refhub.elsevier.com/S1738-5733(18)30540-0/sref1http://refhub.elsevier.com/S1738-5733(18)30540-0/sref1http://refhub.elsevier.com/S1738-5733(18)30540-0/sref1http://refhub.elsevier.com/S1738-5733(18)30540-0/sref1http://refhub.elsevier.com/S1738-5733(18)30540-0/sref1http://refhub.elsevier.com/S1738-5733(18)30540-0/sref2http://refhub.elsevier.com/S1738-5733(18)30540-0/sref2http://refhub.elsevier.com/S1738-5733(18)30540-0/sref2http://refhub.elsevier.com/S1738-5733(18)30540-0/sref3http://refhub.elsevier.com/S1738-5733(18)30540-0/sref3http://refhub.elsevier.com/S1738-5733(18)30540-0/sref3http://refhub.elsevier.com/S1738-5733(18)30540-0/sref3http://refhub.elsevier.com/S1738-5733(18)30540-0/sref3http://refhub.elsevier.com/S1738-5733(18)30540-0/sref4http://refhub.elsevier.com/S1738-5733(18)30540-0/sref4http://refhub.elsevier.com/S1738-5733(18)30540-0/sref4http://refhub.elsevier.com/S1738-5733(18)30540-0/sref4http://refhub.elsevier.com/S1738-5733(18)30540-0/sref5http://refhub.elsevier.com/S1738-5733(18)30540-0/sref5http://refhub.elsevier.com/S1738-5733(18)30540-0/sref5http://refhub.elsevier.com/S1738-5733(18)30540-0/sref6http://refhub.elsevier.com/S1738-5733(18)30540-0/sref6http://refhub.elsevier.com/S1738-5733(18)30540-0/sref6http://refhub.elsevier.com/S1738-5733(18)30540-0/sref6http://refhub.elsevier.com/S1738-5733(18)30540-0/sref7http://refhub.elsevier.com/S1738-5733(18)30540-0/sref7http://refhub.elsevier.com/S1738-5733(18)30540-0/sref7http://refhub.elsevier.com/S1738-5733(18)30540-0/sref7http://refhub.elsevier.com/S1738-5733(18)30540-0/sref8http://refhub.elsevier.com/S1738-5733(18)30540-0/sref8http://refhub.elsevier.com/S1738-5733(18)30540-0/sref8http://refhub.elsevier.com/S1738-5733(18)30540-0/sref8http://refhub.elsevier.com/S1738-5733(18)30540-0/sref8http://refhub.elsevier.com/S1738-5733(18)30540-0/sref9http://refhub.elsevier.com/S1738-5733(18)30540-0/sref9http://refhub.elsevier.com/S1738-5733(18)30540-0/sref9http://refhub.elsevier.com/S1738-5733(18)30540-0/sref9http://refhub.elsevier.com/S1738-5733(18)30540-0/sref9http://refhub.elsevier.com/S1738-5733(18)30540-0/sref10http://refhub.elsevier.com/S1738-5733(18)30540-0/sref10http://refhub.elsevier.com/S1738-5733(18)30540-0/sref10http://refhub.elsevier.com/S1738-5733(18)30540-0/sref11http://refhub.elsevier.com/S1738-5733(18)30540-0/sref11http://refhub.elsevier.com/S1738-5733(18)30540-0/sref11http://refhub.elsevier.com/S1738-5733(18)30540-0/sref11http://refhub.elsevier.com/S1738-5733(18)30540-0/sref12http://refhub.elsevier.com/S1738-5733(18)30540-0/sref12http://refhub.elsevier.com/S1738-5733(18)30540-0/sref12http://refhub.elsevier.com/S1738-5733(18)30540-0/sref13http://refhub.elsevier.com/S1738-5733(18)30540-0/sref13http://refhub.elsevier.com/S1738-5733(18)30540-0/sref13http://refhub.elsevier.com/S1738-5733(18)30540-0/sref13http://refhub.elsevier.com/S1738-5733(18)30540-0/sref13http://refhub.elsevier.com/S1738-5733(18)30540-0/sref14http://refhub.elsevier.com/S1738-5733(18)30540-0/sref14http://refhub.elsevier.com/S1738-5733(18)30540-0/sref14http://refhub.elsevier.com/S1738-5733(18)30540-0/sref14http://refhub.elsevier.com/S1738-5733(18)30540-0/sref14http://refhub.elsevier.com/S1738-5733(18)30540-0/sref15http://refhub.elsevier.com/S1738-5733(18)30540-0/sref15http://refhub.elsevier.com/S1738-5733(18)30540-0/sref16http://refhub.elsevier.com/S1738-5733(18)30540-0/sref16http://refhub.elsevier.com/S1738-5733(18)30540-0/sref16http://refhub.elsevier.com/S1738-5733(18)30540-0/sref17http://refhub.elsevier.com/S1738-5733(18)30540-0/sref18http://refhub.elsevier.com/S1738-5733(18)30540-0/sref18http://refhub.elsevier.com/S1738-5733(18)30540-0/sref19http://refhub.elsevier.com/S1738-5733(18)30540-0/sref19http://refhub.elsevier.com/S1738-5733(18)30540-0/sref20http://refhub.elsevier.com/S1738-5733(18)30540-0/sref20http://refhub.elsevier.com/S1738-5733(18)30540-0/sref20http://refhub.elsevier.com/S1738-5733(18)30540-0/sref20